Inspiration
It is often difficult to gauge a stranger's threat level without any background information. Child abductors, rapists, and fugitives can easily blend in with the general public and inflict harm on the innocent. We wanted to build something to keep the public safe, while also informing law enforcement of criminal activity.
What it does
Our project uses the FBI's database of fugitives from the LA area to inform our users of threats in real-time. RedFlag accesses cameras on wearable devices to inform users of danger when they come face-to-face with a potential threat.
How we built it
First, we used python to upload the FBI's database of fugitives into Microsoft Cognitive Services. Then, we used opencv in C++ to get access to our computer's webcam and output a continuously updated snapshot of the video feed in .jpg format. We sent this file to another python script which accessed Microsoft Cognitive Services' Face API to check for faces in the snapshot. If there is a face in the frame, we check if the face matches one from the FBI database. When a criminal is detected in the frame, we print the criminal's name, and the current location of our user is added to the criminal's profile through the Microsoft Face API. The video feed is also altered to have a canny edge filter when the criminal is in the frame.
Challenges we ran into
Initially, we wanted to use the Hololens to implement this project, but lack of access and difficulty in installing the development tools deterred us from it. We also had a hard time trying to make GET requests to the Microsoft Face API using C++ which is why we switched to python.
Accomplishments that we're proud of
We are proud of the fact that we were able to use Microsoft's facial recognition software in a meaningful manner. Moreover, we are especially happy to have met and collaborated with so many interesting and creative people.
What we learned
We learned how to effectively use API's that we are unfamiliar with. We also learned how to decipher programming languages that we had not encountered before.
What's next for RedFlag
Our project is capable of being exported to much smaller hardware. We used the webcam due to lack of resources. Many wearables like Spectacles, Hololens, and Google Glass have the ability to record from a first person perspective. We would love to develop our project on such platforms in the future. We would also like to incorporate more data on child abductors, rapists, and other dangerous individuals.
Log in or sign up for Devpost to join the conversation.